Tests validate accuracy of face recognition software’s for boarding flights

“We ran simulations to characterise a system that is doing two jobs: identifying passengers at the gate and recording their exit for immigration,” said Patrick Grother, a NIST computer scientist and one of the authors of the resulting report.

“We found that accuracy varies across algorithms, but that modern algorithms generally perform better. If airlines use the more accurate ones, passengers can board many flights with no errors.”

The research team evaluated 29 algorithms designed to perform the so-called one-to-many matching at airports, namely if the person in a photo has a match in a large database of individuals expected to be present.

In reality only a few hundred passengers board a given flight. But the NIST also looked at whether the technology could be viable elsewhere in the airport, especially at security where perhaps 100 times more people might be expected.

Among the report’s findings were the algorithms performed with high accuracy across all variations of gender, nationality and ethnicity. Though false negatives, meaning the software fails to match two images of the same person, were slightly more common for women, they were rare in all cases.